AI Mentors for Student Projects: Spotting Early Issues in Computer Science Proposals
Project Overview
The document highlights the use of a generative AI system, specifically GPT-4o, in enhancing project-based learning (PBL) for computer science students by evaluating their project proposals. This AI-driven approach aims to identify students who may need additional support, thereby improving their educational experience. Preliminary studies indicate that the system is viewed positively by students, aiding in proposal writing and boosting learning motivation. Nonetheless, the success of this initiative hinges on the ability to establish reliable indicators of student success and motivation, as well as addressing challenges such as ambiguity in project proposals and the financial implications of utilizing generative AI. Overall, the findings suggest that while generative AI holds promise in educational settings, careful implementation and consideration of these factors are essential for maximizing its effectiveness.
Key Applications
A software system for collecting and evaluating project proposals to determine student readiness for project-based learning.
Context: Used in high school-level computer science classes for students developing interactive web applications or video games.
Implementation: Implemented as a React.js and Firebase-based web application collecting project proposals and aptitude information.
Outcomes: Engaged users' intrinsic motivation; 88.8% expressed interest in using the system for skill development, and 91.6% wanted to use it for designing motivating project ideas.
Challenges: Vague skills and project descriptions hinder accurate evaluation; the need for reliable grading criteria and predictions of student success.
Implementation Barriers
Technical Barrier
Vague implementation details and skills in project proposals make it hard to evaluate student readiness accurately.
Proposed Solutions: Develop a skill classification rubric for clearer evaluations and improve the system to help students identify specific skills.
Equity Barrier
Costs associated with using generative AI, like API calls, raise concerns about accessibility and equity in education. Explore the use of alternative generative models, including open-source options, to mitigate costs.
Proposed Solutions: Explore the use of alternative generative models, including open-source options, to mitigate costs.
Project Team
Gati Aher
Researcher
Robin Schmucker
Researcher
Tom Mitchell
Researcher
Zachary C. Lipton
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Gati Aher, Robin Schmucker, Tom Mitchell, Zachary C. Lipton
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai